[support]

Hi,

The BSP cannot override those settings, however the BSP can define embedded frameworks that can be referenced by the project and the frameworks may in turn define extra files and flags.

In fact, our BSP-related code is open-source, so you may find some useful examples here: https://github.com/sysprogs/BSPTools/

[support]

Hi,

The dongle does not seem to be very popular, so we won’t be adding a special plugin for it in the near future. However if there is a gdb server that supports it, you can configure VisualGDB to use it by selecting the “Custom GDB Stub” in the debug method selector. If that does not work, feel free to post the problem description here and we will try to help you get it to work.

[support]

Hi,

This looks like a connection problem. Please double-check that all the JTAG connections (TDI, TDO, TMS, TCK) match the layout used with the EVB board. Also try resetting the board manually, but ensure that the reset is disconnected from the JTAG reset (xt-ocd has a bug with the reset on FT2232-based programmers). Finally try lowering the JTAG rate to 1 MHz or even lower.

[support]

Hi,

The easiest way to do both would be to remove all files from solution explorer and re-import them via right-click->Add->Import Folder Recursively. There you can also enable an option to bypass the Source Files/Header Files folders.

[support]

Hi,

We have just rechecked the latest toolchain with another Cortex M3 MCU (STM32F100VB) and could not reproduce the problem, so the fault is not caused by ARM vs Thumb problem.

In order to pinpoint the problem, please use the call stack window together with the disassembly view to find a specific instruction that causes the fault. If this is not helpful, open the Disassembly view before stepping into __libc_init_array() and then step into it, doing one instruction at a time until the fault occurs. If the fault happens on a memory accessing instruction, please check the address used in that instruction. If it is outside the FLASH/RAM ranges, something got wrong with the memory map (e.g. a bug in the linker script or a wrong address somewhere).

[support]

Hi,

Unfortunately, compiling individual files is not yet supported. VisualGDB uses GNU Make as the build system and in this mode Visual Studio can only pass “build project”, “rebuild project” and “clean project” commands, but not individual file commands.

[support]

Hi,

Please try locating the file that defines the UtilInit() function. If it is a C file, ensure that the declaration in the .h file (or the #include<> directive including the .h file) is wrapped with extern “C” {}.

[support]

Hi,

VisualGDB 4.3 does not support the latest STM32 BSP that is required to use STM32CubeMX.

After you upgraded to 5.0 please upgrade your STM32 BSP via Tools->Embedded Tools Manager and then try following the tutorial from the beginning. This should get all paths to work.

[support]

Hi,

Texane ST-Link is separate from OpenOCD, however it’s also buggier. You can try using it, however you may need to unplug and replug your ST-Link if it starts behaving strangely.

Can you verify the lost variables problem with the following program?

int g_Global = 123;
int main()
{
    if (g_Global != 123)
        asm("bkpt 255"); 
}

Does it trigger the breakpoint? What is the address of g_Global?

[support]

Hi,

Are you using the latest VisualGDB 5.0 and the latest STM32 BSP? Does the normal LEDBlink (HAL) project build correctly?

[support]

Hi,

OpenOCD usually takes a few months to support the latest ST devices. Once it supports it, we will update our package.

Regarding the problem with the variables, could you please provide an example? Could you also verify that the variable address is within the RAM region according to the datasheet and that the value is actually lost, not just seems to be lost via the debugger (e.g. output it to a COM port from your firmware).

[support]

Hi,

Yes, simply contact our support with your license key and they will reset your activation.

[support]

Hi,

Good to know you got it to work. Let us know if you encounter further problems.

[support]

Hi,

The “propagate to the environment” flag is only supported for the per-user variables, not for variables set from commands. In the case you mentioned, specifying GitRevision=$(GitRevision) explicitly is the recommended way to go.

[support]

Please double-check that your custom variable has the “propagate to the environment” flag. This can be also caused by a bug in the .Net framework that changes all environment variables to lowercase. You can try $(git_revision) instead. If this does not help, you can explicitly specify it in GNU Make arguments: GIT_REVISION=$(GIT_REVISION). Then VisualGDB will expand $(GIT_REVISION) and pass the actual value to the Make command line.

[Author]

Posts